Oct 13, 2025 · 43m · startup-ideas

My AI Videos Hit 1M+ Views (Veo3 + Sora 2 Demo)

PJ Ace · 31m spoken Greg Isenberg · 6m spoken Ad Voice / Character · 24s spoken Ad Voice / Interjection · 1s spoken Video Ad Narrator / Character · 0s spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

In this masterclass episode of the Startup Ideas Podcast, host Greg Isenberg and viral creator PJ Ace break down the end-to-end production pipeline, creative frameworks, and agency business models behind multi-million-view AI video commercials.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Greg holds 17.4% of the talking time here. How this is scored →

Greg as informed peer 3.0 Guest teaching 5.9 Guest disagreement 0.4 Greg pushing back 0.0
05100:0015:0030:002:51–7:26 · Greg as informed peer 2/10 Origin Financial AI Commercial Audio Playback PJ breaks down his agency's comedic approach to AI commercials to mitigate consumer backlash. Greg facilitates smoothly, asking how novices can bridge the gap from dabbling in AI video to producing viral commercial results.7:30–12:07 · Greg as informed peer 5/10 Step 1: Scripting, Public Domain IP, and Satirical Tropes PJ explains using ChatGPT to brainstorm historical bad financial advice. Greg adds value by synthesizing the workflow into three actionable creative frameworks: public domain IP, juxtaposition, and internet-native hooks.12:09–18:25 · Greg as informed peer 3/10 Step 2: Generating Image Boards and Refining Prompts in Rev PJ demonstrates using Rev to generate iterative image variations before animating. Greg validates the shift from direct text-to-video to image-to-video pipelines, noting how foolish direct generation seems in retrospect.18:28–20:30 · Greg as informed peer 1/10 Team Collaboration and Figma Storyboarding PJ walks through Figma storyboards for high-budget commercial campaigns, detailing the division of labor between writers, directors, and AI cinematographers.20:37–26:50 · Greg as informed peer 3/10 Step 3: Animating Frames with Veo 3 and Competing Models PJ compares video generation models like Veo 3, Kling, and Luma Labs, discussing character dialogue and corporate indemnification. Greg asks about prompt complexity, leading PJ to explain that conversational prompts work just as well as technical camera terms.26:51–30:47 · Greg as informed peer 3/10 Editing, Audio Design, and Outpainting Animation Hacks PJ shares specific post-production stacks and reveals an outpainting hack using picture-in-picture image uploads in Veo 3. Greg expresses amazement at the technique's cost-efficiency.30:48–34:33 · Greg as informed peer 2/10 Ramp Commercial Audio: Horror-Themed B2B Spot PJ walks through the horror-themed Ramp commercial and shares why they discarded their initial text-to-video draft. Greg and PJ discuss using face enhancers to eliminate uncanny AI perfection.34:36–41:31 · Greg as informed peer 5/10 Sora 2 Disruption, Agency Evolution, and IP Licensing PJ and Greg explore how Sora 2 will compress production timelines and create new business models around licensing legacy IP. Greg references recent industry news and frames IP acquisition as an undervalued investment.2:51–7:26 · Guest teaching 5/10 Origin Financial AI Commercial Audio Playback PJ breaks down his agency's comedic approach to AI commercials to mitigate consumer backlash. Greg facilitates smoothly, asking how novices can bridge the gap from dabbling in AI video to producing viral commercial results.7:30–12:07 · Guest teaching 5/10 Step 1: Scripting, Public Domain IP, and Satirical Tropes PJ explains using ChatGPT to brainstorm historical bad financial advice. Greg adds value by synthesizing the workflow into three actionable creative frameworks: public domain IP, juxtaposition, and internet-native hooks.12:09–18:25 · Guest teaching 6/10 Step 2: Generating Image Boards and Refining Prompts in Rev PJ demonstrates using Rev to generate iterative image variations before animating. Greg validates the shift from direct text-to-video to image-to-video pipelines, noting how foolish direct generation seems in retrospect.18:28–20:30 · Guest teaching 6/10 Team Collaboration and Figma Storyboarding PJ walks through Figma storyboards for high-budget commercial campaigns, detailing the division of labor between writers, directors, and AI cinematographers.20:37–26:50 · Guest teaching 7/10 Step 3: Animating Frames with Veo 3 and Competing Models PJ compares video generation models like Veo 3, Kling, and Luma Labs, discussing character dialogue and corporate indemnification. Greg asks about prompt complexity, leading PJ to explain that conversational prompts work just as well as technical camera terms.26:51–30:47 · Guest teaching 7/10 Editing, Audio Design, and Outpainting Animation Hacks PJ shares specific post-production stacks and reveals an outpainting hack using picture-in-picture image uploads in Veo 3. Greg expresses amazement at the technique's cost-efficiency.30:48–34:33 · Guest teaching 5/10 Ramp Commercial Audio: Horror-Themed B2B Spot PJ walks through the horror-themed Ramp commercial and shares why they discarded their initial text-to-video draft. Greg and PJ discuss using face enhancers to eliminate uncanny AI perfection.34:36–41:31 · Guest teaching 6/10 Sora 2 Disruption, Agency Evolution, and IP Licensing PJ and Greg explore how Sora 2 will compress production timelines and create new business models around licensing legacy IP. Greg references recent industry news and frames IP acquisition as an undervalued investment.2:51–7:26 · Guest disagreement 1/10 Origin Financial AI Commercial Audio Playback PJ breaks down his agency's comedic approach to AI commercials to mitigate consumer backlash. Greg facilitates smoothly, asking how novices can bridge the gap from dabbling in AI video to producing viral commercial results.7:30–12:07 · Guest disagreement 0/10 Step 1: Scripting, Public Domain IP, and Satirical Tropes PJ explains using ChatGPT to brainstorm historical bad financial advice. Greg adds value by synthesizing the workflow into three actionable creative frameworks: public domain IP, juxtaposition, and internet-native hooks.12:09–18:25 · Guest disagreement 1/10 Step 2: Generating Image Boards and Refining Prompts in Rev PJ demonstrates using Rev to generate iterative image variations before animating. Greg validates the shift from direct text-to-video to image-to-video pipelines, noting how foolish direct generation seems in retrospect.18:28–20:30 · Guest disagreement 0/10 Team Collaboration and Figma Storyboarding PJ walks through Figma storyboards for high-budget commercial campaigns, detailing the division of labor between writers, directors, and AI cinematographers.20:37–26:50 · Guest disagreement 0/10 Step 3: Animating Frames with Veo 3 and Competing Models PJ compares video generation models like Veo 3, Kling, and Luma Labs, discussing character dialogue and corporate indemnification. Greg asks about prompt complexity, leading PJ to explain that conversational prompts work just as well as technical camera terms.26:51–30:47 · Guest disagreement 0/10 Editing, Audio Design, and Outpainting Animation Hacks PJ shares specific post-production stacks and reveals an outpainting hack using picture-in-picture image uploads in Veo 3. Greg expresses amazement at the technique's cost-efficiency.30:48–34:33 · Guest disagreement 0/10 Ramp Commercial Audio: Horror-Themed B2B Spot PJ walks through the horror-themed Ramp commercial and shares why they discarded their initial text-to-video draft. Greg and PJ discuss using face enhancers to eliminate uncanny AI perfection.34:36–41:31 · Guest disagreement 1/10 Sora 2 Disruption, Agency Evolution, and IP Licensing PJ and Greg explore how Sora 2 will compress production timelines and create new business models around licensing legacy IP. Greg references recent industry news and frames IP acquisition as an undervalued investment.2:51–7:26 · Greg pushing back 0/10 Origin Financial AI Commercial Audio Playback PJ breaks down his agency's comedic approach to AI commercials to mitigate consumer backlash. Greg facilitates smoothly, asking how novices can bridge the gap from dabbling in AI video to producing viral commercial results.7:30–12:07 · Greg pushing back 0/10 Step 1: Scripting, Public Domain IP, and Satirical Tropes PJ explains using ChatGPT to brainstorm historical bad financial advice. Greg adds value by synthesizing the workflow into three actionable creative frameworks: public domain IP, juxtaposition, and internet-native hooks.12:09–18:25 · Greg pushing back 0/10 Step 2: Generating Image Boards and Refining Prompts in Rev PJ demonstrates using Rev to generate iterative image variations before animating. Greg validates the shift from direct text-to-video to image-to-video pipelines, noting how foolish direct generation seems in retrospect.18:28–20:30 · Greg pushing back 0/10 Team Collaboration and Figma Storyboarding PJ walks through Figma storyboards for high-budget commercial campaigns, detailing the division of labor between writers, directors, and AI cinematographers.20:37–26:50 · Greg pushing back 0/10 Step 3: Animating Frames with Veo 3 and Competing Models PJ compares video generation models like Veo 3, Kling, and Luma Labs, discussing character dialogue and corporate indemnification. Greg asks about prompt complexity, leading PJ to explain that conversational prompts work just as well as technical camera terms.26:51–30:47 · Greg pushing back 0/10 Editing, Audio Design, and Outpainting Animation Hacks PJ shares specific post-production stacks and reveals an outpainting hack using picture-in-picture image uploads in Veo 3. Greg expresses amazement at the technique's cost-efficiency.30:48–34:33 · Greg pushing back 0/10 Ramp Commercial Audio: Horror-Themed B2B Spot PJ walks through the horror-themed Ramp commercial and shares why they discarded their initial text-to-video draft. Greg and PJ discuss using face enhancers to eliminate uncanny AI perfection.34:36–41:31 · Greg pushing back 0/10 Sora 2 Disruption, Agency Evolution, and IP Licensing PJ and Greg explore how Sora 2 will compress production timelines and create new business models around licensing legacy IP. Greg references recent industry news and frames IP acquisition as an undervalued investment.

speaking balance: gold is Greg, purple is the guest (3 minute bins)

0:00 · Greg 54.4% · guest 45.6%0:00 · Greg 54.4% · guest 45.6%3:00 · Greg 16.3% · guest 83.7%3:00 · Greg 16.3% · guest 83.7%6:00 · Greg 0% · guest 100%6:00 · Greg 0% · guest 100%9:00 · Greg 38.3% · guest 61.7%9:00 · Greg 38.3% · guest 61.7%12:00 · Greg 0.3% · guest 99.7%12:00 · Greg 0.3% · guest 99.7%15:00 · Greg 12% · guest 88%15:00 · Greg 12% · guest 88%18:00 · Greg 1.1% · guest 98.9%18:00 · Greg 1.1% · guest 98.9%21:00 · Greg 9.9% · guest 90.1%21:00 · Greg 9.9% · guest 90.1%24:00 · Greg 5.2% · guest 94.8%24:00 · Greg 5.2% · guest 94.8%27:00 · Greg 18.9% · guest 81.1%27:00 · Greg 18.9% · guest 81.1%30:00 · Greg 5.5% · guest 94.5%30:00 · Greg 5.5% · guest 94.5%33:00 · Greg 9% · guest 91%33:00 · Greg 9% · guest 91%36:00 · Greg 45% · guest 55%36:00 · Greg 45% · guest 55%39:00 · Greg 24% · guest 76%39:00 · Greg 24% · guest 76%42:00 · Greg 26.1% · guest 73.9%42:00 · Greg 26.1% · guest 73.9%
Sharpest disagreement ▶ 39:02 Calling out OpenAI's launch strategy

PJ casually criticizes OpenAI for using a dirty playbook of removing guardrails to reach top app store ranks before heavily restricting generation.

Hardest push from Greg ▶ 36:26 Ideas remain the true bottleneck

Greg reframes PJ's disruption narrative by arguing that tool commoditization means distinct creative ideas become the ultimate competitive advantage.

Biggest teaching moment ▶ 28:12 Veo 3 picture-in-picture outpainting tutorial

PJ explains an advanced undocumented hack of embedding reference thumbnails into source frames to steer Veo 3's camera reveal.

Greg holds their own ▶ 9:59 Three-pillar virality framework

Greg distills PJ's ad-hoc brainstorming into three clear structural principles for audience retention: public domain IP, juxtaposition, and meme-native relevance.

the scores for every segment, with the reasoning behind each
ChapterTopicGreg as informed peerGuest teachingGuest disagreementGreg pushing backWhy
Origin Financial AI Commercial Audio Playback 2510 PJ breaks down his agency's comedic approach to AI commercials to mitigate consumer backlash. Greg facilitates smoothly, asking how novices can bridge the gap from dabbling in AI video to producing viral commercial results.
Step 1: Scripting, Public Domain IP, and Satirical Tropes 5500 PJ explains using ChatGPT to brainstorm historical bad financial advice. Greg adds value by synthesizing the workflow into three actionable creative frameworks: public domain IP, juxtaposition, and internet-native hooks.
Step 2: Generating Image Boards and Refining Prompts in Rev 3610 PJ demonstrates using Rev to generate iterative image variations before animating. Greg validates the shift from direct text-to-video to image-to-video pipelines, noting how foolish direct generation seems in retrospect.
Team Collaboration and Figma Storyboarding 1600 PJ walks through Figma storyboards for high-budget commercial campaigns, detailing the division of labor between writers, directors, and AI cinematographers.
Step 3: Animating Frames with Veo 3 and Competing Models 3700 PJ compares video generation models like Veo 3, Kling, and Luma Labs, discussing character dialogue and corporate indemnification. Greg asks about prompt complexity, leading PJ to explain that conversational prompts work just as well as technical camera terms.
Editing, Audio Design, and Outpainting Animation Hacks 3700 PJ shares specific post-production stacks and reveals an outpainting hack using picture-in-picture image uploads in Veo 3. Greg expresses amazement at the technique's cost-efficiency.
Ramp Commercial Audio: Horror-Themed B2B Spot 2500 PJ walks through the horror-themed Ramp commercial and shares why they discarded their initial text-to-video draft. Greg and PJ discuss using face enhancers to eliminate uncanny AI perfection.
Sora 2 Disruption, Agency Evolution, and IP Licensing 5610 PJ and Greg explore how Sora 2 will compress production timelines and create new business models around licensing legacy IP. Greg references recent industry news and frames IP acquisition as an undervalued investment.

Statements from this episode (15)

Assertion Not checkable as stated
PJ Ace scaled an AI video agency to 7 figures in months
“Our entire end to end process that we've used to scale to a six figure, seven figure agency in like a few months.”
PJ Ace Oct 13, 2025 ▶ 1:12
Insight
Comedy-first ads mitigate consumer backlash against AI-generated commercial videos
“This is the only way we've found to like mitigate a lot of the kind of like AI disruptions, changing everything, customer pushback. And so, like you said, if people are laughing kind of the whole time or they're entertained, one, they're more likely to watch t…”
PJ Ace Oct 13, 2025 ▶ 4:23
Insight
Effective AI video commercials pair recognizable IP with foreign concepts
“You know, you have to start with like the big idea for the spot. So the big idea for the spot is we like recognizable IP to start so that people have the familiar with the foreign.”
PJ Ace Oct 13, 2025 ▶ 7:30
Insight
Image-to-video AI workflows are cheaper and faster than direct text-to-video
“We like to work with images on the basis that it's a lot easier and cheaper and faster to generate images for the commercial than it is to do everything text to video.”
PJ Ace Oct 13, 2025 ▶ 12:43
Assertion Supported
Genre AI's David Beckham AI commercial hit 230 million views
“Like this IM-A David Beckham ad that we did that got, like, two hundred thirty million views”
PJ Ace Oct 13, 2025 ▶ 19:31
Opinion
Google Veo 3 is currently the best AI model for talking characters
“VO three is probably the best model at the moment for Making characters talk. It's just really realistic. The motion's great.”
PJ Ace Oct 13, 2025 ▶ 20:49
Opinion
Google excels at foundational AI but struggles to build seamless application layers
“Google's, they have great, always foundational technology, but The real struggle for them, I think, always is this application layer that sits on top, is how do we make it seamless and cohesive”
PJ Ace Oct 13, 2025 ▶ 24:20
Insight
All-in-one AI generation platforms are significantly more expensive due to API markups
“Now the downside to using an all in one platform for image generation and video generation is they're using API pricing. So it's just going to be a lot more expensive than if you were to go in”
PJ Ace Oct 13, 2025 ▶ 25:15
Insight
Veo 3 and Sora cannot generate consistent background music across clips
“The reason you wouldn't want VO three or even Sora to like generate a music track is because it's going to generate a new music track for each clip, because it's only Giving you a piece of the pie.”
PJ Ace Oct 13, 2025 ▶ 27:11
Insight
Creators use picture-in-picture starting frames to direct Veo 3 camera pans
“There's weird hacks like that that you can do where you can put picture in picture to describe what you're wanting to like pan the camera over to, or if it was like a script Like a, if they didn't know what the Titanic looked like, you could put a picture of t…”
PJ Ace Oct 13, 2025 ▶ 29:08
Disclosure
Genre AI scrapped its first text-to-video ad due to facial morphing
“Initially we did it all in actually text to video, and we didn't like the look. This is when image to video kind of first came out. So we did the ad in text to video, And it just looked too like AI, like the faces were morphe and it looked like shit. So we act…”
PJ Ace Oct 13, 2025 ▶ 31:54
Assertion Supported
Enhancer.ai allows creators to add realistic skin textures to AI characters
“You can even run a second pass through a program called enhancer.ai where you can even add like acne and details to her face.”
PJ Ace Oct 13, 2025 ▶ 34:02
Prediction Not checkable as stated
Sora will shrink six-week agency video productions to under one week
“This is very disruptive to us because it essentially makes our six to eight week timeline down to what, like a week or less once you're able to kind of tweak and the quality gets better, which it will over the next three to six months.”
PJ Ace Oct 13, 2025 ▶ 35:23
Prediction Not checkable as stated
AI will cut animated TV episode production budgets to $30,000
“I think the budgets will get down to, like, to do another, like, He-Man or Thundercats, you could probably do it for, like, 30 grand on the super lowest end between all the roles if you're doing it, like, at scale and at volume.”
PJ Ace Oct 13, 2025 ▶ 41:11
Prediction Held up
Multi-step AI video editing workflows will soon be baked into Sora
“I do think that, like, for everything I laid out, it's important for now, but it's gonna be baked into Sora's kind of suite, where you can edit clips, you can kind of double tap into this, and this entire process is gonna get, like, a lot faster and a lot chea…”
PJ Ace Oct 13, 2025 ▶ 42:42
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 300 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.